Search

Privacy is one of the essential pillars for the widespread adoption of blockchains, but public blockchains are transparent by nature. Modern analytics techniques can easily subdue the pseudonymity feature of a blockchain user. Some applications have been able to provide practical privacy protections using privacy-preserving cryptography techniques. However, malicious actors have abused them illicitly, discouraging honest actors from using privacy-preserving applications as "mixing" user interactions and funds with anonymous bad actors, causing compliance and regulatory concerns. In this paper, we propose a framework that balances privacy-preserving features by establishing a regulatory and compliant framework called Selective De-Anonymization (SeDe). The adoption of this framework allows privacy-preserving applications on blockchains to de-anonymize illicit transactions by recursive traversal of subgraphs of linked transactions. Our technique achieves this without leaving de-anonymization decisions or control in the hands of a single entity but distributing it among multiple entities while holding them accountable for their respective actions. To instantiate, our framework uses threshold encryption schemes and Zero-Knowledge Proofs (ZKPs).

We propose a middleware solution designed to facilitate seamless integration of privacy using zero-knowledge proofs within various multi-chain protocols, encompassing domains such as DeFi, gaming, social networks, DAOs, e-commerce, and the metaverse. Our design achieves two divergent goals. zkFi aims to preserve consumer privacy while achieving regulation compliance through zero-knowledge proofs. These ends are simultaneously achievable. zkFi protocol is designed to function as a plug-and-play solution, offering developers the flexibility to handle transactional assets while abstracting away the complexities associated with zero-knowledge proofs. Notably, specific expertise in zero-knowledge proofs (ZKP) is optional, attributed to zkFi's modular approach and software development kit (SDK) availability.

Liquidity providers (LPs) on decentralized exchanges (DEXs) can protect themselves from adverse selection risk by updating their positions more frequently. However, repositioning is costly, because LPs have to pay gas fees for each update. We analyze the causal relation between repositioning and liquidity concentration around the market price, using the entry of blockchain scaling solutions, Arbitrum and Polygon, as our instruments. Lower gas fees on scaling solutions allow LPs to update more frequently than on Ethereum. Our results demonstrate that higher repositioning intensity and precision lead to greater liquidity concentration, which benefits small trades by reducing their slippage.

On November 22nd 2022, the lending platform AAVE v2 (on Ethereum) incurred bad debt resulting from a major liquidation event involving a single user who had borrowed close to \$40M of CRV tokens using USDC as collateral. This incident has prompted the Aave community to consider changes to its liquidation threshold, and limitations on the number of illiquid coins that can be borrowed on the platform. In this paper, we argue that the bad debt incurred by AAVE was not due to excess volatility in CRV/USDC price activity on that day, but rather a fundamental flaw in the liquidation logic which triggered a toxic liquidation spiral on the platform. We note that this flaw, which is shared by a number of major DeFi lending markets, can be easily overcome with simple changes to the incentives driving liquidations. We claim that halting all liquidations once a user's loan-to-value (LTV) ratio surpasses a certain threshold value can prevent future toxic liquidation spirals and offer substantial improvement in the bad debt that a lending market can expect to incur. Furthermore, we strongly argue that protocols should enact dynamic liquidation incentives and closing factor policies moving forward for optimal management of protocol risk.

We assess the market risk of the DeFi lending protocols using a multi-asset agent-based model to simulate ensembles of users subject to price-driven liquidation risk. Our multi-asset methodology shows that the protocol's systemic risk is small under stress and that enough collateral is always present to underwrite active loans. Our simulations use a wide variety of historical data to model market volatility and run the agent-based simulation to show that even if all the assets like ETH, BTC and MATIC increase their hourly volatility by more than ten times, the protocol carries less than 0.1\% default risk given suggested protocol parameter values for liquidation loan-to-value ratio and liquidation incentives.

Decoding strategies play a crucial role in natural language generation systems. They are usually designed and evaluated in open-ended text-only tasks, and it is not clear how different strategies handle the numerous challenges that goal-oriented multimodal systems face (such as grounding and informativeness). To answer this question, we compare a wide variety of different decoding strategies and hyper-parameter configurations in a Visual Dialogue referential game. Although none of them successfully balance lexical richness, accuracy in the task, and visual grounding, our in-depth analysis allows us to highlight the strengths and weaknesses of each decoding strategy. We believe our findings and suggestions may serve as a starting point for designing more effective decoding algorithms that handle the challenges of Visual Dialogue tasks., Comment: Accepted at INLG 2022